Return Styles: Pseud0ch, Terminal, Valhalla, NES, Geocities, Blue Moon.

Pages: 1-

Conways Game of Pixels

Name: Anonymous 2011-09-07 5:54

It can run turing machines, imagine a LISP machine inside that, simulating some computation.

Name: Anonymous 2011-09-07 6:05

I wonder if hashlife could give performance similar to native code.

Name: Anonymous 2011-09-07 6:32

you're ignorant

Name: Anonymous 2011-09-07 6:54

>>1
Now imagine that the computation was some physics engine simulating a self-aware being and its environment.

Name: Anonymous 2011-09-07 7:46

JACKSON 5 GET

Name: Anonymous 2011-09-07 7:50

>>2
First problem is aliasing/alignment artifacts. That is, even if you have lots of semantically similar components, unless they are also located in the same position in respect to the containing quadtree nodes they are going to hash into a polynomial number of different keys. Plus the same problem for temporal alignment of processes, if I understand correctly.

Then, any computation in the Game of Life is by necessity local, I mean, you don't have anything like random access, your signals have to physically propagate across the virtual space. So it's not very efficient model to begin with, similar to Turing Machine, actually. Hashlife might speed up this, but only if you have larger exactly repeating patterns in the entire region where the "wires" in question are located. I think. I'm not sure.

All in all I suspect that you will lose nothing except some problems if you go for "hash lambda calculus" or "hash SKI" instead.

Name: Anonymous 2011-09-07 8:35

Oh lispers. I know that you alwats want to be slower than turtle, but this is a new low even for you.

Name: Anonymous 2011-09-07 14:46

>>6
your signals have to physically propagate across the virtual space.
Just like real laws of physics. Not that anyone claimed most CAs as efficient (although some might be possible to implement efficiently in hardware).

Name: Anonymous 2011-09-07 21:36

The best way of doing physics simulation is doing them in real life, you're not limited by computational power or lack thereof.

Name: Anonymous 2011-09-08 1:37

>>9
You cannot always inspect the exact state of the system at all times, nor even the full properties of some particle. You can observer by interact as you are part of the physical system. Simulations are useful because you can statically inspect the system with maximum precision/detail you would like.

Name: Anonymous 2011-09-08 2:03

game of dicks

Name: Anonymous 2011-09-08 5:08

>>10
You can observer by interact
(โ—”โ—กโ—”)

Name: Anonymous 2011-09-08 6:20

>>11
fuck, you beat me to it and you also got dubz

Don't change these.
Name: Email:
Entire Thread Thread List